Notes and Presentation from the HUG talk on Nov 18th - integrating solr & katta for scaled reliable search

Replies to This Discussion

Permalink Reply by Jason Venner on December 1, 2009 at 7:51am

There are 2 solr config files in the setup.
We have the solrconfig.xml that is bundled as part of each index, and we have a solrconfig that is used by each front solr instance.

The way this works is that each solr index is hosted by a katta node, and handles a special type of request the KattaRequest. Each of these indexes is a named shard.

In addition to these embedded solr instances, there are 1 or more front end solr instances. These instances provide a user interface, I commonly run them in jetty, using a hacked version of the bundled examples.
These solr instances take the incoming requests and distributed the queries to the specified shard names. The results are collected, merged and returned to the requesting user/agent.

• ▶ Reply

Permalink Reply by Jason Venner on December 3, 2009 at 3:38pm

Configuration files
front-end-solr-config.xml is the solconfig for the solar instances that serve user queries
zoo.cfg is the configuration for my zookeeper cluster
katta.zk.properties is the katta configuration that points to the zoo.cfg nodes
katta.master.properties is the katta file that sets the distribution policy for the index shards
back-end-solr-config.xml is the solrconfig for my individual shards.

Attachments:

• katta.zk.properties, 1 KB
• katta.master.properties, 1 KB
• backend-solr-config.xml, 36 KB

• ▶ Reply

Permalink Reply by Jason Venner on December 3, 2009 at 3:38pm

more files

Attachments:

• zoo.cfg, 1 KB
• front-end-solr-config.xml, 38 KB

• ▶ Reply

▶ Reply to Discussion